DPBRN learned several lessons regarding its administrative structure in its early years; several aspects evolved accordingly. For example, although our PSC was the main decision-making body, we soon realized that we needed a Regional Directors Committee (RDC) to decide how best to implement those decisions. We next learned that the RDC could easily get its agenda bogged down by discussing details of particular studies, so next we formed study implementation groups, whose only agenda item was to discuss a single study. However, we next learned that this same team really needed to shepherd the study in each phase, from study design, to pilot testing, to field phase management, to data analysis and manuscript preparation. Thus we formed Study Teams for each study (led by the Study PI). Weekly meetings between the National Network Director group and the Data Coordinating Center group fostered communication between these two groups, adjusted priorities weekly, and established clarity about who was supposed to do what and when. We later realized that our Regional Coordinators (research assistants) needed to discuss agenda items not specific to a particular study, so we began Regional Coordinator meetings. We also learned that we needed a formal document to guide administration of publications and presentations, completing in 2006 the DPBRN publications policy [22], which in 2009 served as the template for the CONDOR three-network policy [23]. Coincident with all this we saw the communication value of taking minutes at each meeting, with a focus on recording Decisions, assigning Action Items with deadlines, and holding persons accountable for those items. Although this amounted to a significant number of committees, it improved our effectiveness substantially. Because the network must have many 'balls in the air' at the same time to maximize its productivity, we believe that this evolution in administrative structure fostered our productivity in the 2005-2012 cycle. Therefore, we propose the same administrative structure for NDPBRN, depicted in the graphic below. The best structure is the one that leads to the best function and performance. We believe that structure is best compared across networks not by comparing assertions about structure or planned collaborations, but by using objective measures of output (i.e., the ones delineated in Section B.I., p. 222).